A distributed data-stream architecture finds application in sensor networks for monitoring
environment and activities. In such a network, large numbers of sensors deliver continuous
data to a central server. The rate at which the data is sampled at each sensor affects the
communication resource and the computational load at the central server. In this paper, we
propose a novel adaptive sampling technique where the sampling rate at each sensor
adapts to the streaming-data characteristics. Our approach employs a Kalman-Filter (KF) …

According to the vision of Mark Weiser, that the physical world will be connected with
pervasive networks and that everyday devices will be able to sense their relationship to
humans and to each other, the ability to sense a broad set of physical phenomena, will
become a common aspect of small, embedded devices. Furthermore such devices shall
communicate with each other to organize and coordinate their actions. The physical world
presents an incredibly rich set of input modalities, including acoustics, image, motion …